Thank you for raising this important question Benjamin. I have read the two publications. The review is an interesting one and agrees with my own personal experience with direct injection into tendons – initial good response but recurrence is common, though it is hard to know if recurrences are due to re-injury consequent on return to normal activities during the early ‘honeymoon stage’ when the presence of corticosteroid is still present within the tissues. This problem is hard to address because the rapid and dramatic relief commonly experienced following corticosteroid injection (CSI) has a profound effect on behaviour, especially in athletes and workers whose jobs require significant loading. It is those folk who commonly get painful tendinopathy and they are keen or pressured to return to training or work who commonly receive CSI. Personally I am quite adamant that there must be at least three weeks of ‘relative rest’ post CSI and in the case of glenohumeral joint CSI for Stage I & II adhesive capsulitis, the advice is at least a month – regardless of (even especially with) an excellent reduction of pain initially. There seems to be no doubt that CSI is the only current treatment for adhesive capsulitis that gives a very high proportion of initial rapid and better than 90% recovery, but the recurrence problem and the legitimate concerns with structural weakening of tendon structure makes it imperative that we search for a reliable alternative to CSI for painful tendinopathy.

The study by Coombes BK et al is an excellent one that does not suffer from the common problem of treating physiotherapy as a ‘black box’ to gloss over the fact that ‘physiotherapy’ has often been portrayed as hot packs, ultrasound and other passive therapies. The authors describe the physiotherapy which conformed to an evidence based use of manual therapy and exercise. I have graphed out their results to give a visual impression of their results and I have attached the Excel file for other readers to view and consider. It is impelling.

There is a clear benefit for CSI for lateral epicondylitis in the short term (4 weeks post CSI with or without physiotherapy), compared to placebo with or without physiotherapy. However, placebo injection with or without physiotherapy is clearly superior by 26 weeks, and all patients receiving placebo and physiotherapy had an excellent clinical outcome by 52 weeks. The question I ask is: “is the recovery by 52 weeks in the placebo group related to the passage of time, or a direct effect of the intervention(s)?” My own clinical experience is that left alone, the great majority of cases of lateral epicondylitis recover by 2 years after onset. The patients in the study had a median duration of symptoms of 16 weeks (range 6 weeks – 4 years). Obviously the patient with 4 years of ongoing trouble has failed natural history, but those with symptoms lasting 6-18 months prior to study enrolment are likely to benefit from natural history and this would pump up the good results from placebo with or without physiotherapy. I would like to see a study of clearly defined cases of isolated lateral epicondylitis stratified according to symptom duration where placebo injection and physiotherapy (as described in the current study or a close equivalent) are compared. Also I would like to see a similar symptom duration-stratified study comparing placebo with something else like dextrose that is known to cause fibrous tissue proliferation, rather than the opposite effect that CSI appears to produce.

The authors of the lateral epicondylitis study correctly point out that while there is no statistical benefit evident from physiotherapy, it must be pointed out that rapid pain relief in the short term was noted from manual therapy, and that perhaps 7-8% more patients achieved an excellent clinical result if physiotherapy is added to placebo. Perhaps there is a (small) subgroup for whom manual therapy and exercise is the key valuable ingredient?

The other point I would like to make is that while direct CSI to painful tendons is clearly not valuable in the medium and long term, there will be cases of severe and disabling pain where the treatment is justifiable in the short term – simply as a matter of pain relief. My advice in this circumstance is that should CSI be recommended, a very long period of 2 months or more of rest and supervised gradual return to loading (exercise and graduated return to work / training) after 1-2 months of rest be considered an essential part of the treatment package. Hard to implement I know, but I think that this advice and ongoing rehabilitation for several months probably are now essential. Another study required!

The message here is that we must not throw the baby out with the bathwater!

Further, I would like to stress that while CSI to tendons should now be considered a less than optimum treatment, the same cannot be said of CSI to non-tendinous structures like joints. For example, at this time, management of adhesive capsulitis of the shoulder depends heavily on intra-articular CSI to the glenohumeral joint in the first two stages of the disease. The clinical benefits of rapid reduction of distressing night pain are many, and the risks appear to be minimal (though in the case of infection, though rare, it is serious).

Just as an aside, my co-workers and I have two papers currently under review with a journal which report on the clinical outcomes of primary care shoulder pain patients. A subgroup were offered CSI into the subacromial bursa, acromioclavicular joint and glenohumeral joint under strict conditions and strongly advised to engage in ‘relative rest for 3 weeks post injection. Those patients receiving the CSI usually responded very well in the first 4 weeks (but not always) and those receiving CSI to the subacromial bursa had a noticeable ‘clinical bounce’ whereby the mean pain and disability values worsened between the 4 week and 12 week follow up. The same ‘clinical bounce’ was not as obvious in those receiving joint injections. Also half of all recurrences were associated with a documented re-injury or noticeable strain event – almost exactly the same as Coombes et al found in their study. I will make these papers available in due course.

thanks for that response, fascinating and thought provoking

I agree with you in that the Coombes study is very good and as you say, it is interesting to consider why people get better over time in this study

I suspect, as you do, that a lot of the improvement is simply the disease natural history and that the CSI is having a negative long term affect on the natural history, probably related to its anti-inflammatory properties which may well worsen the tendinopathic changes relative to a placebo group

Thanks for that snippet of you work, look forward to reading it!

I agree with you on baby and bath water, CSI definitely has a role but I am just a bit more cautious about using it and by understanding the harms a bit better, I hope to avoid causing it in patients!

Certainly it's overall risk/benefit and if you can avoid surgery in a patient with severe pain by buying some time, it can clearly be worth it. I still just use the harms to explain to patients that a second injection is probably not a good idea, as if they had a 2-4 week benefit from the first, if we keep repeating this for short term gains we are probably going to make things worse in the long term!

The short term gain from hitting inflammation and the long term harm does suggest that treatments which unregulated the inflammatory response in the short term may well be the way to go for long term gain?

Benjamin: Not sure exactly what you are meaning in your last paragraph.

I agree that if there is recurrence without evidence of re-injury, then a second injection is not a viable option - except with Stage I & II idiopathic / non-traumatic adhesive capsulitis. I see no harm in up to 3 injections in this case. I don't think there is risk of weakening the capsule-on the contrary, the capsule, by definition is thickened and tight. Without having actually audited my cases, my feeling is that I rarely do even two injections and only once or twice have regarded three as necessary.

There is one issue that I believe is very important in trying to understand the clinical response to steroid injection, and is not often adequately appreciated. Patients not uncommonly report that after steroid injection, there is some improvement but not a large amount, and this improvement slowly emerges over a week or so. It is clear that a steroid injection clearly removed from the painful site often helps symptoms. This was well illustrated in a study where one arm of the sample received CSI to the painful subacromial bursa and the other arm received CSI into the buttock on the opposite side(1). The differences in outcomes for the two groups were not significantly different. Also patients receiving a steroid injection for a specific problem will sometimes report that another problem far removed in anatomical terms also improved. This is clear evidence of a systemic anti-inflammatory effect following CSI. These systemic effects are rarely dramatic compared to the sudden ablation of pain in 1-5 days that are seen regularly when the painful pathology successfully targeted. In the case of a CSI to a specific pathology that results in minimal to moderate and/or slow response, I would argue that the painful tissue has not been infiltrated or that the condition is not actually inflammatory in nature.

Guided CSI improves target specificity, but one has to be very sure that the imaged target is actually the source of pain and that the source of pain is inflammatory. What I am saying is that good steroid injections are not easy to do. The injection technique itself is straightforward enough in most cases in the extremities (spinal targets are a different ball game), but selection of patients, advice to patients in the immediate post injection period out to one month, supervised guidance in return to normal loading are crucial to getting good results. The post injection care is simple enough but needs time in patient education – something which is hard to do consistently in busy community primary care, but is possible. What is not so easy is the diagnostic aspect. Selection of patients and identification of the primary source of pain is much more complex than what we have believed in the past. To emphasize this point I recommend that the recent work by Cadogan et al(2,3,4,5,6) (I am a co-investigator and co-author), has made this very clear I think. We are currently working on another paper looking at the diagnostic accuracy of the clinical examination in relation to guided diagnostic block of the glenohumeral joint. We hope to submit this paper sometime this year.

References

1. Ekeberg OM, Bautz-Holter E, Tveita EK, Juel NG, Kvalheim S, Brox JI. Subacromial ultrasound guided or systemic steroid injection for rotator cuff disease: randomised double blind study. BMJ 2009; 338: a3112-

2. Cadogan A, Laslett M, Hing WA, McNair PJ, Coates MH. A prospective study of shoulder pain in primary care: Prevalence of imaged pathology and response to guided diagnostic blocks. BMC Musculoskelet Disord 2011; 12: 119-

3. Cadogan A, Laslett M, Hing W, McNair P, Williams M. Interexaminer reliability of orthopaedic special tests used in the assessment of shoulder pain. Man Ther 2011; 16: 131-135.

4. Cadogan A, Laslett M, Hing W, McNair P, Taylor S. Clinical predictors of a positive response to guided diagnostic block into the subacromial bursa. J Rehabil Med 2012; 44: 877-884.

5. Cadogan A, McNair P, Laslett M, Hing W. Shoulder pain in primary care: diagnostic accuracy of clinical examination tests for non-traumatic acromioclavicular joint pain. BMC Musculoskelet Disord 2013; 14: 156-

6. Cadogan A, McNair P, Laslett M, Hing W, Taylor S. Diagnostic accuracy of clinical examination features for identifying large rotator cuff tears in primary health care . Journal of Manual and Manipulative Therapy 2013; 21: 148-159.

Dear Benjamin

The tendinopathy (ranging from inflammation to tissue degeneration) tend to be more frequent in the young population, while disruptions (partial or total) are seen in most sports population above 40 years of age.

Conservative treatment offers the advantage of avoiding surgery and its related complications (infection, nerve damage and deltoid).

infiltration of glucorticoid - Acts decreasing inflammation and hence the pain, thereby facilitating rehabilitation. The harmful effects of these substances are well known (atrophy and deterioration of tissue quality, which could prejudice any future surgical repair) and therefore should be used with caution, and recommended maximum three infiltrations, with a minimum interval of three months including

Should be taken after discuss the advantages and disadvantages with the patient

Thanks for those interesting references.

I shall get reading!

In terms of the last paragraph, I was trying to get at the 'good' and 'bad' of inflammatory processes. The tissue changes in tendinopathy generally fit with failing homeostasis, ie collagen disruption and disorganisation, with a failing attempt to heal this damage. As part of this the inflammatory system is upregulated both in terms of cells and cytokines, even if 'cold' degenerative tissue.

In my opinion the short term gains from steroid are part of the same package which consists of the long terms harms as demonstrated by Coombes, ie by switching down inflammatory processes one gets short term pain gain, but long term the effect of this is more tissue damage, consequently greater tissue response to this damage, and consequently poorer long term outcomes. That's my theory anyway.

I find the diagnostic overlap very interesting, a lot can probably be explained by the overlap in innervation between joint, bursa and ACJ.

http://www.ncbi.nlm.nih.gov/pubmed/23429268

I agree that it can be incredibly hard to pinpoint 'sources' of pain. Perhaps seeing one single 'source' is too simplistic for many patients, we see many with multiple peripheral drivers and central drivers. That's why pain is so fascinating I guess.